Insurance Charges Prediction using Machine Learning

This project builds a predictive model to estimate medical insurance charges based on an individual’s characteristics such as age, sex, BMI, number of children, smoking status, and region. The model uses a Random Forest Regressor and is built using Python and common machine learning libraries.

---

Dataset

The dataset used is [insurance.csv], which contains:

• age: Age of the primary beneficiary
• sex: Gender (male/female)
• bmi: Body Mass Index
• children: Number of dependents
• smoker: Smoking status (yes/no)
• region: Residential area (southeast, southwest, northeast, northwest)
• charges: Medical charges billed by health insurance

---

Project Workflow

1. Data Preprocessing

• Loaded the dataset using pandas.
• Checked for null values and data types.
• Performed exploratory data analysis (EDA) to understand distributions.

2. Data Visualization

• Used matplotlib and seaborn to visualize:
  • Age distribution by gender
  • Regional breakdown of gender counts

3. Encoding & Feature Engineering

• Label encoded categorical variables (sex, smoker, region).
• Used Recursive Feature Elimination (RFE) with LinearRegression to select the top 5 features.
• Applied a log transformation to the charges column to reduce skewness and improve regression accuracy.

4. Standardization & Train/Test Split

• Selected features: age, bmi, children, smoker, region.
• Split the data into training and testing sets using train_test_split.
• Standardized the features using StandardScaler.

5. Model Training

• Trained a RandomForestRegressor on the scaled training data.

6. Model Evaluation

Evaluated the model using:

• Mean Squared Error (MSE): ~0.128
• R² Score: ~0.857

7. Making Predictions

• Accepted new user input in DataFrame format.
• Encoded and scaled the input before prediction.
• Outputted predicted insurance charges (in log scale).

---

Libraries Used

• pandas
• numpy
• matplotlib
• seaborn
• sklearn

---